Objective High quality and effective occupational health service (OHS) is heavily dependent on smart use and analysis of data. To provide OHS in an innovative and economically efficient way using real world data and machine learning (ML/AI) we have developed a proof-of-concept (PoC) of an analytical platform HeBA (Health for Business Analytics) to support everyday clinical practice of employee health management.

Methods The innovative approach of HeBA is that it uses naturally collected data of OHS provision and offers analyses based decision support for OHS physicians and for employers. It has modules for physicians (seamlessly integrated with electronic health record system), for individual employees and employer OH administrators. For managing employees’ health in companies HeBA analyses indicators of quality of working life - presentism, stress-level, motivation, sick-leave days and allows to see correlations between working-life quality indicators and health indicators by department, company, economic sector or occupation. Companies can get annually action plans for health management and make an analyses based decisions in employees’ and workplace investments. OH physicians can get rule-based decision support in their everyday practice.

Results HeBA has been tested and used in our OHS clinic serving our clients - 10 000 employees. An example of possibilities: we can measure prevalence of poor self-rated health by occupations in datapool of all employees; make comparisons of poor health between departments in a company and assess association of it with workplace risk-factors and health indicators. Qualitative feedback to the PoC from 3 major user groups - physicians, employer administrators and individual employees - has shown very high satisfaction rate and given guidance for further development potential of HeBA.

Conclusion The proof-of-concept has proven its value in improving the OHS quality and efficiency in Estonia. Further research and development is planned for international validation and introducing ML/AI in decision support solutions.